Is there any research into how many bugs are introduced per line of code in a serious project?
I mean, you could always write a bash script that crawls random github pages, count line numbers in the project, and divide total errors / line count and do statistics on them yourself if you want, and then publish a paper on it.
If you went through the time to do it, I'd read the paper.
>>60074566
It's generally safe to assume the bigger the project the more likely a major bug is to be there.
>>60075345
Yeah but by how much? Is it a uniform linear increase or is it exponential?
How much is the average bug count? and line count?
How much variance or gaps are there between lines of code in general? Or bugs in general?
How much does developer count affect this?
How much does repo popularity affect this?
How much does the line addition / subtraction per day affect this?
What is the aggregate line and bug count, graphed over time?
What is the rate of errors per day of a currently large repo since its birth?
What is the rate of errors per day adjusted with popularity bias of a currently large repo since its birth?
How much does popularity affect known bug counts?
How many bugs are fixed before a bug report is made?
How does popularity increase or decrease that number?
Pretend I repeated all these questions again but asked how programming language affected this.
>>60074566
https://www.fastcompany.com/28121/they-write-right-stuff
>>60074566
Not sure but you would have to normalize by the number of Pajets working on a project to work out the true figure.